The Internet of Things continues to be the shiny object that is driving the electronics industry, from the giants to the start-ups come and many makers. The IoT continues to ride along pivoting with the dominant platform of the smart phone, often a key link in the chain of a complete system, although tablets may be waning.
Drones (officially, Unmanned Aerial Vehicles - UAVs) have proven their value for extremely high-end applications, although even if fully self-driven automobiles are technically feasible, they are not likely to ever hit the public streets. However, the Advanced Driver Assistance Systems (ADAS) that spin off from the exotic research vehicles have shown a real value in everyday cars - at least in luxury models. One hopes that some day similar spin-off products will take root from the chaos and noise that the IoT is instigating.
Action cameras have gained traction, but while there is a great deal of activity in televisions and streaming content delivery, a visible market shift toward the big fancy screen remains elusive.
The power-saving green movement in electronics is still underway but has been elbowed to the back seat again. This is unfortunate because energy conservation can have an immediate and almost universal benefit. Still, the convenience of portability and the inconvenience of recharging batteries that deteriorate will keep power consumption high on the list of desirable features in electronics. Meanwhile, municipal power grids struggle to keep the lights on. In the automotive market, much lower gasoline prices have taken the luster off of electric, hybrid, and similarly green automobiles.
The beginning of the year is a good time to review the dynamics of the electronic industry so that programs close to our hearts can be realigned with the realities of the greater world. It's also a good time to tickle the memory.
IoT not Out
There has been no end to the Internet of Things (IoT) bandwagon, and every CEO and marketer has found a way of hitching on. This will continue through 2015 and 2016. The measure of this mythical market grows with every mention, but getting this grand Internet of Things to all play together is still largely a figment of the imagination.
About the only thing universal about the IoT is the Internet itself and the existence of IP addresses. A variety of networks and protocols can be used to get to the Internet including:
- 3G, 4G, LTE cellular
- numerous sub-GHz wireless schemes
The challenge is that different companies in the same market space make different choices. Each company thinks it has the best idea and that others will surely follow its lead and attach on. The Internet may be utilized for communication, but there is no universal code for "the temperature in Chicago at 2:00 pm on June 14, 2014, in Fahrenheit," let alone "what will it take to get the person in the camera to buy that blouse right now?"
From building controls and home gateways to automotive infotainment systems to the factory floor, interoperability is vital. Today, for any end-point to effectively use information from any other end-point or even a central server, all ends have to be very carefully selected and controlled - and often built and managed by the same vendor. Users of electronic systems are accustomed to this difficulty, but are not happy with it, whether determining operation on the current sugar-coated operating system, realizing that the slowest wireless device in the chain limits the ultimate throughput, or cursing (or enjoying) that if Apple didn't make or certify it then it won't work in the set-up. But the Internet will remain just a fragmented jumble of Things unless a viable mechanism is in place to allow anything on the Internet to easily connect and communicate with anything else.
There are company and committee proposals for standardizing the way devices communicate on the Internet, but there are many rather than just one unifying method. Interoperability among most Things, network nodes, and computers or servers is necessary for the Internet of Things (IoT) to really blossom. But for the most part each OEM seems to be promoting only its own approach.
One exception to the lack of standards is where either the iPhone or iPad or an Android device is utilized. However, that is about as far as commonality goes; it's not as though the Fitbit app will blend its data with a NOAA (US Department of Commerce's National Oceanic and Atmospheric Administration) weather app even if they each talks on the iPhone. But the old school PC has lost its luster as a necessary platform, as has Windows as the OS and the x86 as the primary processor architecture. There isn't even one "Cloud" but many clouds all implemented in their own atmosphere.
The great thing about standards, of course, is there are so many of them. There are numerous companies and groups trying to form their own standard for the IoT including:
- AllSeen Alliance
- IEEE P2413
- Industrial Internet Consortium
- Open Interconnect Consortium
- Thread Group
But the best standard is always the de facto standard that isn't forced on the industry but is chosen by the users because it just plain works, is readily available, and is widely supported.
The presence of these groups have so far had little impact with the most promising being the IEEE committee since it is supposedly fairly unbiased. It will take a long time for a real winner to emerge, if one ever does. While many companies may be promoting IoT standards, none of those companies have any presence or influence in the most explosive part of the IoT - the Things (which were called Embedded Systems for decades).
The plethora of these new Things are unmanned end-points on the Internet, ever-smaller and more-sophisticated than before, possibly just 8-, 16-, and 32-bit microcontrollers (MCUs) monitoring a single sensor or driving an actuator. Maybe they run the smart watch or the body sensors. Any IoT standard that hopes to be successful will have to accommodate the highly constrained requirements of these Things. Security and encryption overlays cannot impair the operation of these real-time-responsive systems nor increase their power consumption, space, performance, or cost. The ARM processor architecture (not the x86) has a dominant and growing presence in these Things, along with many other architectures. Yet the named big companies have no significant market role in traditional real-time embedded systems.
Right now, the Internet of Things is a rally cry around which to pitch a profusion of products. Many are prior products now re-introduced with some new smart phone or Internet connectivity. For example, home controls have been available for over 30 years, as has remote monitoring. However, cheap image sensors were needed to bring remote viewing into the picture. It took the prevalence of always-on broadband in the home before the Internet became the favored medium for monitoring and control rather than 22-gauge telephone lines. But the fundamental home or building controls worked just fine over telephone wire. A/C on. A/C set to 78. Meeting room lights off. Open front gate.
Basically, the Internet is just a huge data pipe into and out of the home, business, and factory. It is difficult to keep up with the classifications of data and the services available over the Internet. The PC was encroached upon by the tablet which is heavily dependent upon the Internet. The telephone is likely to be a Voice-over-Internet Protocol (VoIP) phone if a land line exists at all. Television programs come over-the-air (OTA), from a satellite, on an RF cable, or streamed or over-the-top (OTT) of the Internet. Gaming can be a high-intensity activity not only for the console and user, but for the Internet as well. Home controls including energy management tend to be fairly low data-rate items for the Internet unless constant video feeds are employed. It is surprising that faxes are still used - by doctors and lawyers, of all people! Of all of these, only home controls (or building controls in a commercial setting) would normally be considered Things on the Internet since no human is directly involved.
Once an Internet gateway is part of the home or business, the opportunity expands greatly for "Things" to be linked via the Internet to other Things, getting information from distant sources, passing data to centralized databases, and being utilized by services in "the Cloud". There is a constant wrestle of who has access to, control of, and revenues generated from all that data flow: the cable guy, the phone company, the wireless service provider, a municipal utility, Big Blue, Big Brother, the search engine, or some cloaked hacker.
The Phone/Tablet Platform
The last ten years has seen an overwhelming surge of smart phones, particularly Apple's iPhone and numerous companies' Android-based phones. The smart phone and similar tablets have become a platform in the way that the PC has been a platform - a more powerful processing engine that can greatly extend the utility of an embedded device or a simpler end-point while providing user-interface and display capabilities that could be difficult to build into a standalone embedded system. With access to the Internet through a data channel, the smart phone also can act like a gateway to the Internet and points far away to access other data sources, greater services, storage centers, and the Cloud. The beauty of the smart phone or tablet is its tremendous portability and the millions of people who use them constantly. Most of the new Things being added to the Internet are set up and managed using a smart phone or tablet app.
Flaw of Large Numbers
Of course, what is a market without an enormous growth projection? Numbers tossed around for the Internet of Things multiply like Internet-enabled rabbits, with presenters struggling to come up with ancient-Greek/Latin-based prefixes to sum up the totals. Bigger is always better and any market researcher can add just one more zero to the projected number. Is peta- bigger than zetta-? Comparisons far surpass the hairs on a man's head (Biblical), stars in the sky, grains of sand in the Sahara, drops of water in the ocean, tons of TNT, or transistors on a slab of silicon. Mathematicians are reportedly scrambling to enlarge infinity.
A year ago, Cisco predicted 7.7 zettabytes (1021) of data would be handled by the world's data centers in 2017. http://www.cisco.com/web/tomorrow-starts-here/files/growth_in_the_cloud.pdf In 2011, Cisco predicted that 2020 would see 50 billion Things on the Internet. http://www.cisco.com/web/about/ac79/docs/innov/IoT_IBSG_0411FINAL.pdf
Intel is now using a figure four times that, saying there will 200 billion connections to the Internet by 2020. http://www.intel.com/content/www/us/en/internet-of-things/infographics/guide-to-iot.html
While Cisco may have a share of the network market and Intel dominates PCs and servers, neither company has any part of the smart phone or tablet market or any of the exploding new Things market. But ARM and numerous other microprocessor architectures other than Intel's x86, plus the billions of microcontrollers (MCUs), serve the real-time and embedded systems market place quite well, and this is where the Things come from. Clearly, ARM-based processors are the de facto standard for smart phones and tablets.
There are two issues with such large numbers. Where will the money come from to buy all of those things and once they've been bought, can they all talk, relate, or extrapolate together? Where are the algorithms and computers to make sense of all this data? What information can be pulled out of the data? Today, an enormous amount of data exists documenting activity in the stock market, readily available to computers. And yet the price or even direction of tomorrow's or next month's stock trades is still far from predictable. Therefore, will increasing the data set 10x appreciably improve the analytical output? How much better are weather and market predictions than those found by reading the Farmer's Almanac, the Horoscope, or tea leaves? Or just a little sane judgment?
The second big issue follows "How much better is the analysis?" with "What was the cost of the grand electronic systems that gave this better result?" It's not just the cost of the computers, but cost of the endpoints that gather the raw data (the Things), the communications links to transmit the data, the servers to coalesce the data, and development of the algorithms to extract intelligence from the data. The payback will need to cover all of those costs, initially and continually. A recent presentation proclaimed victory in making a 1% improvement resulting in a $2B savings. As an absolute number $2B is a lot of money, but the company was still burning $198B on jet fuel.
Automobiles and Advanced Driver Assistance Systems
The automobile has been a vehicle for great expansion for the electronics industry since the '80's after improved fuel economy and reduced air pollutants were legislated, and only intelligent electronic systems could provide the solution. Devices that enhanced safety followed, with more-attractive sales features like convenience and drivability further driving the use of electronics in automobiles.
The utopian idea of a driverless car (along with some DARPA temptations) has motivated development of numerous techniques to provide information about other vehicles, road conditions, weather, obstacles, and dangers outside of the vehicle during real-life driving situations. Each of these Advanced Driver Assistance Systems (ADAS) is very useful individually, whether they are watching for unintended lane departures, vehicles or people in blind spots, objects behind a car that is backing up, unexpected objects in the road, insufficient following distance, pedestrian movement, potential intersection crashes, or just a good panoramic view surrounding the vehicle. These ADAS systems are making cars safer to drive even when a human is at the wheel and are attractive enough to justify their cost to the purchaser, especially in the higher-end cars where they first show up.
There may not be consistency in how such systems work across various automobile makers and the assorted systems are still being integrated. However, as more ADAS are installed in cars, feedback is obtained from drivers, systems are improved, cost comes down, and more people can enjoy the benefits as the systems get into more widely-used automobiles.
The most advanced systems, and those most closely tied to IoT, may struggle to get the needed infrastructure in place to be useful. Some proposed systems look for specific real-time information, sensors placed at the roadside and centrally monitored, and data collected from other vehicles in the last 15 minutes. That level of information requires the buy-in of the local municipality (read: tax base), integration with various agencies responsible for the roads involved, maintenance of hardware and upgrade of software - let alone enough automobiles participating to make it all worthwhile. Security and privacy issues regarding the data can also get in the way.
Getting the money together for such infrastructure, equipment, computers, and communications can be a daunting task, especially when pot-holes aren't being filled. The market becomes very small if only new cars from a few models of a couple of carmakers can supply and make use of a system installed along one or two routes - and the information is really only useful during rush hour.
Maybe Driving is the Distraction
The auto industry continues to be in an electronic quandary. Consumer electronics advances on an annual basis, but it takes five to seven years for highly-reliable custom-fit equipment to be designed into production automobiles. Cassette decks and CD players today have almost no value. Just keeping Bluetooth up-to-date in a car is difficult. Navigation systems are placed more predominantly on the dashboard than the speedometer or clock. DVD players and gaming systems have been in cars for a decade. Wi-Fi on wheels is being put in cars to pull down the Internet and stream videos while traveling down the road. Of course, a key to the cell phone's success was the ability to make productive use of the long drive time going to work or back home. Now there are smart phones, some well-blended into the car, bringing an unending array of apps that give access to non-automotive capabilities via the Internet.
Yet, starting the first of 2015, Austin, Texas, joins other cities around the world in greatly restricting the use of cell phones while operating moving vehicles. In Austin, the driver can not touch the cell phone: not to answer a call, not to read a txt message, not to change a music selection, not to shop online, not to pinch to zoom a photo, and certainly not to view who is waiting at the front door and unlocking the house if it's the housekeeper. The driver is supposed to be paying attention to the road, possibly with the help of ADAS. Everything else, including entering navigation addresses, is considered a dangerous distraction - and now an illegal activity. Older cell phones and older or low- to mid-range cars may now be legislated as incompatible.
The difficulty is for the auto makers to embrace fast-moving consumer-oriented technology in a pleasing way while assuring that the driver is still fully alert to driving conditions and in complete control of the vehicle. Costly property damage and human lives are at stake. It is a difficult balance to strike. People are willing to pay premiums for electronics in cars, but regulators are pushing back to keep the public safe.
Go Take a Video, Soon in the Round
A new market seems to have been found by the GoPro camera, allowing adventurists and athletes to capture the excitement of their activities in all the realism of high-definition video, at least from a rather annoying fish-eye perspective (correctable). While other cameras are offered for this market, GoPro seems to be the only one that has captured the imagination and a nice market share in the space. Even though the name doesn't scream "hands-free thrills camera", GoPro is as synonymous with the market space as Kleenex is (in the USA) with nose tissues.
Remarkably, the product has not appreciably added to the piles of videos of disastrous drunken or stupid stunts performed just for attention that have proliferated since the cell phone camera started feeding YouTube. Perhaps the "Pro" in GoPro, the higher price, and careful management of videos promoted by the company have kept the focus on true adventure and spectacular visuals rather than human foolishness. However, it's hard to see this market growing significantly or associated products providing new opportunities for semiconductor volumes. There are not a dearth of budding adventurists on the sidelines just waiting for lower pricing or even higher resolution. But the company stock soared to four times its Initial Public Offering price before approaching the end of 2014 at about 2.75x the IPO. Not bad for a couple of California surfer dudes.
Some new competitive cameras are now coming out that offer an all-around view. These so-called "720" cameras (720° a bit of a misnomer; perhaps "2-axes" would be better) capture moving or still images in all directions at once. The images of two back-to-back full fish-eye cameras, each with >180° field of view, are stitched together, capturing front, back, right, left, and up and down. This is a somewhat unnatural view except to some lizards and fish, but it certainly gives a full visual context of any scene. For the 4K resolution, full motion video, and real-time de-warping, quite a bit of processor horsepower is needed.
The market for such cameras is developing but is likely far smaller than your everyday camera market. There is often little of interest on the ceiling or the floor and it rarely matters what is behind a person that is moving forward. However, for security applications and automotive safety applications, the one camera could replace many image sensors (six and more on some cars). Real-time software will remain intensive as there are more areas to monitor, objects to detect, and events of interest.
Not so Rosy
Not all extreme cameras have been fully embraced by mankind. Nerds wearing spy camera glasses that add a reality to their perhaps unreal world seem to have drawn more ire than admire from the general public. Google has retreated from their experiment of constantly recording and augmenting life with Google Glass. In spite of what Reality TV might lead one to believe, most people don't really want to live their life under the microscope of a camera lens, especially if it's in the hands of someone else. "Too much information" may also apply here. Not everyone is a trained fighter pilot needing swarms of data interjected to their everyday life. The wonders of Hollywood's The Matrix and RoboCop should not always be the blueprint for the next new product (flop). Fun research, but product failure does not spur an economically strong electronics or semiconductor industry.
Throwing TV a Curve
Every year a new angle on television is thrust upon the public. A person only needed one eye open to see that 3-D TV was going to go nowhere, just as the movies have since the late 1950's. 60 Hertz displays help smooth the moving picture, especially in sports and high-action scenes, and can take advantage of Blu-ray features, but going to 120 Hz TVs and beyond is dependent upon creating images that were never there - resulting in little true image enhancement.
From a manufacturing point of view, it is impressive that screens can be made now with a slight concave curve to them, but a curve to a TV screen is more gimmick than image improvement (Fig. 1). If anything, the effect is actually worse when more people are all trying to watch the same screen because the sweet spot is more narrow. The concept is now being pushed on some cell phones as well (Figure 3), but provides even less benefit there. It is hard to see how these curved screens can justify more than a 10% price increase to the end user.
The current high-cost trend in TVs is going to Ultra-High-Definition, UHD, or 4K (Figure 2). These 3,840 x 2,160 pixel images put twice the pixels in each direction of the well-engrained 1080 screens, but to really offer 4x the resolution, the source material has to also be provided at that high resolution. Today, not only is 4K source material a scarce item to find, it requires a huge upgrade in storage, bandwidth, and equipment to capture (the cameras) and transmit to the consumers (over cable, satellite, Internet…) or spool to Digital Video Recorders (DVR).
Of course, every step up in television is hampered by availability of cameras, content, industry and network support, transmission systems, and displays, but the cost of such upgrades seem to be getting beyond what most consumers can justify since their last big purchase that was supposed to be the ultimate. Thus, the uptake of larger screen 4K televisions is likely to crawl along slowly. And there is an 8K standard right behind 4K.
Proof is as near as the Internet. Recently the ultimate TV made by a CE manufacturer that always claims it has the next big thing offered a 105" 4K 120 Hz UHD curved screen 3-D Smart TV for only $120,000. However, two pair of 3-D glasses and the remote control are included. A rewards program can get the buyer 2% back in points and free shipping. However, the model is sold out, so it is not clear when shipping will actually take place. Game, set, match. Fail. http://www.amazon.com/Samsung-UN105S9-Curved-105-Inch-Ultra/dp/B00L403O8U/ref=cm_cr_pr_pb_t
Growth, but with Some Stammering
Everything thrown at the Internet will not survive, but the IoT is definitely the next stage of the Internet and the build up is well underway. Making increased IoT growth explode will require all the new Things to pass information to all the other Things on the Internet too, and that may take some time to materialize. Just how many devices will attach to the Internet, only time will tell. The real benefit of the self-driving car dream is all the ADAS that may be thrown off from the dream car. Specialized cameras will contribute some to new applications, but the motivation for upgrading to the fanciest TVs is weak right now, awaiting infrastructure and content.